Isolation and Characterization of Proteins Associated with Histone H3 Tails in Vivo*

The histone H3 amino-terminal tails play an important role in regulating chromatin transcription. Although the mechanisms by which the H3 tail modulates transcription are not well understood, recent discoveries of specific interactions of regulatory factors with H3 tails suggest that H3 tails are a key player in the precise regulation of transcription activity. To investigate the recruitment-based action of H3 tails in chromatin transcription, we purified H3 tail-associated proteins from HeLa cells that stably express epitope-tagged H3 tails. This approach resulted in the identification of multiple histone methyltransferase activities and transcription regulatory factors that are specifically associated with expressed H3 tail domains. Point mutations of Lys-9 and Lys-27 to block cellular modifications of the tail domains completely abolished the association of specific factors, including HP1 and several repressors. Importantly, our transcription analysis revealed that the purified factors can significantly stimulate p300-mediated transcription from chromatin templates. These results implicate that the H3 tail, when accessible in relaxed chromatin, acts as a transcriptional regulator by mediating recruitment of specific sets of cofactors.

The nucleosome is a basic unit of chromatin in eukaryotic cells, consisting of an octamer of four core histones wrapped by 146 bp of DNA (1)(2)(3). The most dynamic parts of the nucleosome are amino-terminal domains of core histones that are directly involved in transcriptional regulation (4). The most important feature of these tail domains is their reversible modifications by acetylation, methylation, phosphorylation, and ubiquitination, among others (5). The complexity and dynamic character of histone tails and their modifications present a challenge for detailed studies on how DNA within chromatin structure could be accessible to regulatory factors. Noting that the H3-H4 tetramers organize the central part of the nucleosome, whereas the H2A-H2B dimers organize the more peripheral part of nucleosomes, H3-H4 tails appear to play a major role in transcriptional regulation (6). Indeed, in recent transcription analysis using recombinant chromatin templates with selected tail mutations, it was shown that the H3 and H4 tails and their modifications are required for transcriptional regulation, whereas the H2A and H2B tails are dispensable (7). Although these results confirm the importance of H3 and H4 tails, the underlying mechanism(s) involved in transcriptional regulation by these tail domains still remain to be elucidated.
H3 and H4 tails have been shown to be an important determinant of internucleosomal interactions within chromatin, which are critical for regulating chromatin compaction and structure (8,9). However, in addition histone tails are found to act as a protein interaction module to facilitate the recruitment of factors that will establish a specific functional domain in chromatin. It is also evident that distinct patterns of histone modifications play a critical role in binding of particular chromatin-regulating factors to specific chromatin domains. This selectivity of histone modifications led to the "histone code" hypothesis (10 -12), which proposes that histone modifications serve as a cognate mark for the recruitment of specific proteins to specify unique downstream functions. For example, methylations of H3 at Lys-9 and Lys-27 have been linked to transcriptional repression, because of their recognition by heterochromatin protein HP1 (heterochromatin protein 1;) and polycomb, respectively (13)(14)(15)(16)(17). In contrast, methylation of H3 at Lys-4 creates binding sites for WDR5, which is a component of MLL1/2 and Set1 HMT 3 complexes, and Chd1, which is the chromodomain-containing protein in the SAGA and SLIK complexes (18,19). Recent studies also demonstrated that specific acetylations of H3 and H4 create marks for the binding of bromodomains present in many transcriptional regulators such as GCN5, PCAF, and TAFII250 (20). Thus H3 and H4 tails should be the initial player to accommodate chromatin remodeling signals during gene activation.
To define the specific roles of H3 tails in the transcription process, we have established HeLa cell lines stably expressing * This work was supported in part by the Margaret E. Early Medical Research Trust. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. epitope-tagged H3 tails for the purification of H3 tail-interacting proteins. Our results underscored the interaction of H3 tails with multiple cellular components with activities that significantly boost p300-mediated transcription from chromatin templates, supporting the role of H3 tails as a key factor in establishing a specific transcriptional environment. Importantly, blocking repressive methylations of Lys-9 and Lys-27 of expressed H3 tails significantly antagonizes the association of known repressive factors, confirming these modifications act as major regulatory marks in the recruitment process.

EXPERIMENTAL PROCEDURES
Plasmids-cDNA sequence encoding amino acids 1-40 of human H3 was PCR-amplified by use of a 5Ј primer (5Ј-ATTG-CGGCCGCATGCATATGGCTCGTACTAAACAG-3Ј), which introduced NotI and NsiI sites, and a 3Ј primer (5Ј-TAAGAATT-CTCTGCAGGTGAGGCTTTTTCACACC-3Ј), which introduced EcoRI and PstI sites. The generated products were digested with NotI-EcoRI and inserted into NotI-EcoRI-linearized pBS SK ϩ (pBS-1xnH3). The plasmid (pBS-2xnH3) containing two copies of H3 tail cDNA was generated by inserting the NsiI-EcoRI-digested H3 tail cDNA fragment into PstI-EcoRI-digested pBS-1xnH3. Note that NsiI and PstI produce compatible cohesive ends. To yield the plasmid (pBS-4xnH3) containing four copies of H3 tail cDNA, two copies of H3 tail cDNA digested from pBS-2xnH3 were ligated into PstI-EcoRIdigested pBS-2xnH3. Finally, pBS-8xnH3 was generated by inserting four copies of H3 tail cDNA into PstI-EcoRI-digested pBS-4xnH3. After subcloning eight tandem repeats of the synthesized H3 tail cDNA in pBS, the inserted DNA was excised with NotI and EcoRI and ligated into NotI and EcoRI sites of pIRES containing FLAG and HA tags to generate the plasmid (pFHnH3-IRESneo) for mammalian expression. The same procedure was used to construct the plasmid encoding mutant H3-K9/27R tail domain except that mutations in Lys-9 and Lys-27 to arginine were introduced to original H3 cDNA by using the QuikChange II site-directed mutagenesis kit (Stratagene).
Purification and Identification of H3 Tail-interacting Proteins-HeLa cells were transfected with pFHnH3-IRESneo (wild type and K9/27R) using Lipofectamine (Invitrogen) and selected with G418 (500 g/ml) for 2 weeks. G418-resistent colonies stably expressing H3 tails were grown in spinner culture in Dulbecco's modified Eagle's-phosphate (Irvine Scientific) supplemented with 10% bovine calf serum. Nuclear extracts were prepared as described (21). For purification of H3 tail-interacting proteins, nuclear extracts (300 mg) were fractionated through a phosphocellulose P11 column (Whatman). The P11 BC1000 and BC1500 fractions containing expressed H3 tails were dialyzed against BC300 and applied to M2-agarose affinity chromatography (Sigma). After extensive washings with BC300 containing 0.1% Nonidet P-40, H3 tail-associated proteins were eluted from M2-agarose by using FLAG peptide (200 ng/l). Expression and purification of ectopic H3 tails were confirmed by Western blot using both FLAG (Sigma) and HA (Santa Cruz Biotechnology) antibodies. The purified H3 tail-associated proteins were analyzed by data-dependent tandem mass spectrometry using an LCQ DecaXP ion trap mass spectrometer (Thermo Finnigan) as described previously (22). The mass spectral data were searched against human protein sequence data base using SEQUEST (23). Antibodies employed in Western blot analysis were as follows: histone H1, histone H2B, and CARM1 antibodies were obtained from Upstate; HMGB1 antibody was from Abcam; TIF1␤, nucleolin, and Spt16 antibodies were from Santa Cruz Biotechnology; Lamin-A/C antibody was from Sigma; ␣-tubulin antibody was from Cell Signaling Technology; PARP1, HP1, and G9a antibodies were kind gifts from Drs. Comai, Rice, and Stallcup, respectively; and antibodies against TRAP150 and SAP130 were kindly provided by Drs. Roeder and Martinez.
In Vitro Transcription Assay-Chromatin templates were assembled as described (25) by using recombinant histones and the chromatin assembly factors ATP-utilizing chromatin assembly and remodeling factor and NAP1 (nucleosome assembly protein 1). The plasmid DNA template containing adenovirus major late core promoter and p53-response elements was as described (26). FLAG-tagged p53 was expressed in bacteria and purified as described recently (26). FLAGtagged p300 protein was expressed in insect Sf9 cells and purified on M2-agarose according to standard procedure. Transcription assays were performed using p53 (15 ng), p300 (20 ng), and acetyl-CoA (10 M) as reported previously (27) except that H3 tail-associated factors were added together with p300 and acetyl-CoA (Fig. 5A).

Purification of H3 Tail-associated Factors from HeLa-derived
Cell Lines-To gain mechanistic insight into recruitmentbased action of H3 tails in transcription, we generated HeLa cell lines that stably express double (FLAG and HA)-tagged histone H3 tails for the purification of tail-interacting proteins by a combination of conventional and immunoaffinity chromatographies. Specifically, the N-terminal tail cDNA sequence encoding residues 1-40 of human H3 was PCR-amplified and inserted, as eight tandem repeats, into a mammalian expression vector that has been modified to include the FLAG and HA epitope coding sequences (Fig. 1A). Amino acids 39 -43 of H3 pass between the gyres of the DNA superhelix through chan-nels formed by the minor grooves of the DNA terminus and the central turn near the dyad axis (6); hence we employed residues 1-40 of H3 excluding other parts of H3 embedded within the nucleosome particles. This approach allows us to identify proteins capable of binding specifically to H3 tails in vivo. Because a single copy of the tail was highly inefficient for our detection and purification (data not shown), we put eight copies of the tails for more efficient/concentrated preparation of the associated proteins. The tail fragments span the regions encoding a nuclear localization signal (28), thus allowing accumulation of expressed tails in the HeLa nucleus (Fig. 1A). Indeed our Western analysis confirmed that major fractions of expressed tails were transported from the cytoplasm into the nucleus (Fig. 1B).
In addition to wild type H3 tails, we also expressed H3 tails with specific mutations at Lys-9 and Lys-27 to block their cellular methylations. Methylations of Lys-9 and Lys-27 are known to be repressive marks (29) that could antagonize the association of active factors with H3 tails. Therefore, mutations at Lys-9 and Lys-27 will cause the release of proteins that recognize these repressive modifications, allowing effective purification of proteins specifically recognizing active tail modifications. To obtain highly purified tail-interacting proteins from the nuclear extract, we first fractionated the extract on a phosphocellulose P11 column. We found that the H3 tail-associated proteins were efficiently eluted in 1.0 and 1.5 M KCl fractions (P11 1.0/1.5) (Fig. 1A). Pools of P11 1.0/1.5 fractions were further purified by M2-agarose affinity chromatography under stringent conditions (300 mM KCl, 0.1% Nonidet P-40) (Fig.  1A). Analysis of the H3 tail-associated proteins on SDS-PAGE revealed 13 major bands that copurified with ectopic H3 tails and were not detectable with normal HeLa nuclear extract ( Fig. 2A). Taken together, these results document the success of purification of H3 tail-interacting proteins directly from cells.
Identification of H3 Tail-associated Factors-After a two-step purification of tail-associated proteins, proteins purified with wild type H3 tails were subjected to mass spectrometry analysis to identify individual components. Results are summarized in Fig. 2A (lane 2). Implying the chromatin-specific activity of proteins, our analysis revealed the presence of histone methyltransferases (ASH1, CARM1, EuHMT, G9a, and MLL3), histone demethylase JMJD2C (jumonji domain protein 2C), and histone deacetylases HDAC5 and HDAC9 in the H3 tail-associated proteins. We also detected a stable association of nucleolin and FACT (Spt16 and SSRP1 (Spt16 is suppressor of Ty 16 homolog; SSRP1 is structurespecific recognition protein 1)) within the purified tail-associated proteins. Because FACT and nucleolin both have been shown to assist transient release and redeposition of H2A-H2B dimers from nucleosome during transcription (30,31), our results raise the possibility that H3 tails play a role in initial association of these activities to facilitate transcription through the nucleosome. Our results also indicated the association of several factors that have been linked to transcriptional regulation (SAP130, SAP145, SAP155, Mi-2b, TRAP150, and TIF1␤). Notably, the identification of SAP130, SAP145, and SAP155, which are key subunits of splicing factor SF3b, as components of the purified H3 tail-associated proteins, support the possibility of functional linkage between the chromatin remodeling process and cotranscriptional splicing (32,33). As expected from recent studies, we also identified the association of three histone-binding proteins (HP1␣, HP1␤, and HP1␥) to expressed H3 tails. We assume that these interactions are established by cellular di-and trimethylation of Lys-9 of expressed H3 tails (see below). Our data also revealed the association of the H3 tail with HMGB1, H1, and PARP1 that have been shown to interact with nucleosomes (34 -36). Because it has been proposed that HMGB1, H1, and PARP1 can compete with each other for their binding to nucleosomes (36,37), H3 tails localized at the entry and exit points of nucleosomal DNA could function as a regulator for their stable association to nucleosomes. Authenticity of the proteins identified by mass spectrometry was confirmed by immunoblot using available antibodies (Fig. 2B, lane 2). Taken together, we conclude that H3 tail domains expressed outside of chromatin can act as a binding motif for recruitment of multiple proteins that can regulate chromatin competency. Purification of H3 tail-associated proteins. A, schematic diagram of purification of H3 tail-associated proteins. "A," "M," and "P" indicate sites of acetylation, methylation, and phosphorylation of H3 tails, respectively. Nuclear extracts from HeLa-derived H3 tail-expressing cells were fractionated over P11 ion exchange column and step-eluted with BC buffer as indicated. A pool of 1.0 and 1.5 M KCl eluates was subjected to M2-agarose affinity purification. Control purification was conducted in parallel with similar fractions derived from HeLa cells. Western blotting was performed with anti-HA and anti-FLAG monoclonal antibodies as indicated. Lane 1, mock-purified fraction from control HeLa nuclear extract (NE); lane 2, proteins purified with wild type H3 tails; lane 3, proteins purified with H3 tails mutated at Lys-9 and Lys-27. B, nuclear localization of ectopic H3 tails. Nuclear and cytoplasmic extracts were prepared and analyzed by Western blotting with FLAG antibody. ␣-Tubulin and lamin-A/C were used as markers for cytoplasmic and nuclear fractions, respectively. Lanes 1 and 3, wild type H3 tails; lanes 2 and 4, K9/27R mutated H3 tails.

Cellular Modifications of Ectopic H3 Tails-Because very lit-
tle is known about prominent histone modifying activities in HeLa cells, we determined whether ectopically expressed H3 tails are subjected to any cellular modifications by Western blot analysis. We first checked the acetylation status of the H3 tails with antibodies specifically recognizing each of four acetylation sites (Lys-9, Lys-14, Lys-18, and Lys-23) in the H3 tails. We detected acetylation of ectopic H3 tails at all four lysine substrates (Fig. 3A, lane 2). We next investigated methylation of the H3 tails using antibodies specific to methylated Lys-4, Lys-9, Lys-27, and Lys-36 of H3 (Fig. 3D, lanes 2, 5, and 8).
Because the ⑀-amino group of lysine residues may be mono-, di-, or trimethylated, we examined all possible methylation statuses. Our analysis revealed di-and trimethylation of Lys-4 in ectopically expressed H3 tails, which are linked to transcriptional activation, but the same analysis showed no detectable monomethylation of Lys-4. We also observed significant levels of di-and trimethylation of Lys-9 and monomethylation of Lys-27, which are generally correlated with transcriptional repression. Similar analysis using antibodies specific to methylated Lys-36 also confirmed mono-and trimethylation at Lys-36, the latter of which is linked to active transcription. In further analyses, we also detected steady-state levels of phosphorylation at Ser-10 (Fig. 3B, lane 2) and dimethylation of Arg-17 (Fig. 3C, lane 2). Thus these results clearly indicate that specific patterns of modifications can be introduced onto free H3 tails within living cells.
Histone Modifying Activities of H3 Tail-associated Proteins-Because we found that H3 tails are associated with factors that could modulate the chromatin state, we next determined histone modifying activities present in the H3 tail-associated proteins. As shown in Fig. 4A, the purified tail-associated proteins contain stimulatory activities to preferentially methylate H3 in equimolar mixtures of all four core histones (lane 2). To determine the substrate specificity of HMT activities in the tail-associated proteins, H3 methylation was further characterized by Western blot analysis using highly specific antibodies that discriminate mono-, di-, and trimethylation states of Lys-4, Lys-9, Lys-27, and Lys-36 (Fig. 4C, lanes 2,  5, and 8). Consistent with the mass spectrometry results indicating the stable association of HMTs for Lys-4 and Lys-9 methylations (Fig.  2), our analysis with antibodies recognizing methylations of Lys-4 and Lys-9 confirmed the presence of HMT activities specific for di-and trimethylation of Lys-4 and monoand dimethylation of Lys-9. Parallel experiments with antibodies specific for methylated Lys-27 and Lys-36 revealed the presence of HMT activities capable of monomethylation of Lys-27 and trimethylation of Lys-36. These results are somewhat surprising because no known HMT (e.g. EZH2) for Lys-27 methylation could be detected in our mass spectrometry results (see Fig. 2A), and human HMT specific for Lys- 36 has not yet been discovered. Therefore, future studies should identify enzymatic activities for these modifications in the H3 tail-associated proteins. It is also possible that a certain HMT shows broader specificities when associated with other factors within the purified tail-associated proteins.
Consistent with the presence of CARM1 in the purified tailinteracting proteins, additional analysis showed dimethylation of Arg-17 by the purified proteins (Fig. 4B, lane 2). Interestingly, although expressed H3 tails are acetylated at all major lysine substrates and phosphorylated at Ser-10 (Fig. 3, A and B), we could not detect any histone acetyltransferase and kinase activities in the purified proteins (data not shown). The absence of these activities argues for the dynamic action of histone acetyltransferase and kinase via their highly transient interactions with ectopic H3 tails. Taken together, these results show that H3 tails can act as a binding domain for multiple HMT activities. Purification and Identification of Mutant H3 Tail-associated Proteins-The above results revealed multiple cellular modifications of ectopic H3 tails (Fig. 3), raising the possibility that each of the observed modifications may influence the association of specific proteins directly or indirectly. Some of the observed modifications are generally associated with active chromatin states (e.g. acetylations at Lys-9, Lys-14, Lys-18, and Lys-23 and methylations at Lys-4 and Lys-36), whereas other observed modifications are generally associated with repressed chromatin states (e.g. methylations at Lys-9 and Lys-27). To assess the contributions of only activating H3 tail modifications on protein binding, we mutagenized Lys-9 and Lys-27 of the ectopic H3 tail to block their cellular methylations. Using this approach, we could determine whether disruption of these repressive methylation signals in ectopic H3 tail would antagonize association of any specific factors that could have negative effects on transcription. Although Lys-9 can be either methylated or acetylated, recent mass spectrometry analysis identified abundant dimethylation at Lys-9 in the major population of H3 (38); thus we assumed that our mutation of Lys-9 will mainly block cellular methylation of Lys-9.
We first checked the modification status of the expressed mutant H3 tails. As expected, Western analysis revealed that mutations at Lys-9 and Lys-27 did not affect acetylations of expressed H3 tails at Lys-14, Lys-18, and Lys-23 (Fig. 3A, lane  3), but it completely abolished methylation at Lys-9 and Lys-27 ( Fig. 3D, lanes 3, 6, and 9). Our analysis also confirmed remarkable effects of mutations on dimethylation at Lys-4 and monomethylation at Lys-36, but little or no effect on trimethylations at Lys-4 and Lys-36 (Fig. 3D, lanes 3, 6, and 9) and phosphorylation at Ser-10 (Fig. 3B, lane 3). Unexpectedly, a similar analysis with antibody specific for dimethylated Arg-17 (Fig.  3C, lane 3) revealed that mutations at Lys-9 and Lys-27 significantly reduced Arg-17 methylation. Notably, Western analysis with the antibody for acetylated Lys-9 generated a false-positive result (even with Lys-9 mutated to Arg) (Fig. 3A, lane 3), indicating that the antibody does not solely react with acetyl-lysine epitope but also interacts with other parts or modification of the H3 tails.
To identify the proteins associated with mutant H3 tails, the purified proteins were again subjected to mass spectrometry analysis ( Fig. 2A, lane 3). Significantly, our results confirmed the absence of all three HP1s (␣, ␤, and ␥) that are known to be associated with Lys-9-methylated H3 in the mutant tail-associated proteins. In a more surprising finding, mutations of ectopic H3 tails also abolished the binding of Mi-2b, HDAC5, HDAC9, and TIF1␤, which have been shown to play roles in transcription repression (39 -41), and two uncharacterized proteins (FLJ12800, FLJ13639). However, mass spectrometry analysis could identify within the mutant H3 tail-associated proteins all other factors copurified with the wild type H3 tails ( Fig. 2A, compare lane 3 with lane 2). Again, the mass spectrometry results were confirmed by immunoblot analysis with available antibodies (Fig. 2B, lane 3). Interestingly, although G9a and CARM1 were shown to be associated with both wild  . HMT activities of H3 tail-associated proteins. A, histone methylation by H3 tail-associated proteins. The H3 tail-associated proteins were subjected to HMT assays with recombinant core histones and [ 3 H]AdoMet (SAM indicates AdoMet) as described recently (42). The samples were analyzed by 15% SDS-PAGE and fluorography. Lane 1, mock-purified control; lane 2, proteins purified with wild type H3 tails; lane 3, proteins purified with H3 tails mutated at Lys-9 and Lys-27. B, arginine-specific methylation by H3 tailassociated proteins. HMT assays were identical to A except that recombinant H3 protein and unlabeled AdoMet were used for each reaction. Western analysis was performed with antibody specific for methylated Arg-17 of H3 as described in C. Lane 1, mock-purified control; lane 2, proteins purified with wild type H3 tails; lane 3, proteins purified with H3 tails mutated at Lys-9 and Lys-27. C, lysine-specific methylation by H3 tail-associated factors. HMT assays were identical to B. Western analysis was performed with antibodies specific for methylated Lys-4, Lys-9, Lys-27, and Lys-36 of H3 as described in Fig. 3D. Lanes 1, 4, and 7, mock-purified control; lanes 2, 5, and 8, proteins purified with wild type H3 tails; lanes 3, 6, and 9, proteins purified with H3 tails mutated at Lys-9 and Lys-27.
type and mutant H3 tails by mass spectrometry, our Western blot analysis demonstrated that the mutations significantly inhibit association of G9a and CARM1 with ectopic H3 tail (Fig.  2B). Consistent with these results, the proteins associated with mutant H3 tails showed reduced HMT activities for mono-and dimethylation of Lys-9, which probably were caused by G9a (Fig. 4C, lanes 3 and 6), and dimethylation of Arg-17, which was by CARM1 (Fig. 4B, lane 3). Moreover, although monomethylation of Lys-27 was slightly decreased (Fig. 4C, lane 3), other HMT activities for di-and trimethylation of Lys-4 and trimethylation of Lys-36 (Fig. 4C, lanes 6 and 9) were not significantly affected by these mutations. Therefore, our results provide direct evidence that methylations at Lys-9 and Lys-27 play an important role in the interaction of H3 tails with specific factors, supporting that highly selective recognition/retention of proteins by a certain modification mark the H3 tail domain. These results also confirm the feasibility of our approach to purify H3 tail-associated factors and to identify factors associated with specific H3 tail modifications, setting the stage for more detailed analysis with differentially mutated tails in our future studies.
Effects of H3 Tail-associated Proteins on Chromatin Transcription-To characterize the function of the H3 tail-associated proteins in transcription, we next assessed the ability of the tail-interacting proteins to regulate transcription from chromatin or DNA templates. Because regulatory activities known for both transcriptional activation and repression were purified by using ectopic H3 tails, we were particularly interested in examining selective action of the H3 tail-associated proteins in p300-mediated chromatin transcription. Transcription assays with recombinant chromatin templates containing p53-response elements upstream of core promoter sequences were carried out as described previously (26), except that the H3 tail-associated proteins were added together with p300 and acetyl-CoA (Fig. 5A). As previously reported (7,42), transcription from chromatin template was completely dependent upon p53, p300, and acetyl-CoA (Fig. 5B, lanes 5-8), whereas DNA transcription showed a dependence only on the activator p53 (Fig. 5B, lanes 1-4). Significantly, addition of the H3 tail-associated proteins resulted in a distinct boost of p300-mediated transcription from chromatin template (Fig.  5C, lanes 13-15), whereas similar experiments in the absence of p300 and acetyl-CoA showed no effect of the H3 tail-associated proteins in transcription (lanes 16 -18). In parallel experiments with free DNA, H3 tail-associated proteins showed no effect in transcription (Fig. 5C,  lanes 1-6). These results suggest that the effect of H3 tail-associated proteins depends on the prior action of p300 on chromatin template. Because blocking repressive methylations at Lys-9 and Lys-27 inhibited the binding of several repressors (HP1, G9a, and TIF1␤) to expressed H3 tails (Fig. 2), we next checked if the mutant tail-associated proteins differentially contributed to the observed transcriptional effects. Interestingly, the proteins purified with mutant tails stimulated transcription similarly to that observed with the wild type tail-associated proteins (Fig.  5C, lanes 19 -21). Although HP1 and G9a have been implicated in activation as well as repression (43,44), it is important to note that our analysis was restricted to acetylation-mediated transcription without including any HMT and AdoMet. Therefore, our future studies will determine whether isolated proteins differentially contribute to transcription depending on the methylation state of chromatin template.

Selective Recognition of H3 Tail by Regulatory Factors-Previous experiments performed with recombinant chromatin
templates have demonstrated the requirement of H3 tails in transcription of relaxed chromatin templates (7,42); thus, comparing the level of transcription from intact chromatin versus H3 tailless chromatin shows a significant repression of transcription by deletion of the H3 tail. Although these results bear an important implication on the critical role of H3 tails in chromatin function, how H3 tails regulate transcription is unclear. According to a widely accepted model, H3 tails can serve as physical interaction surfaces for the recruitment of specific regulatory factors to modulate downstream transcription activities (4, 10 -12, 45). In fact, to screen proteins capable of binding to the H3 tail, recent studies applied affinity columns prepared with either unmodified or specifically modified H3 tail peptides (46 -49). These peptide affinity columns have successfully been used to identify the tail-binding proteins from HeLa nuclear extract. However, a major drawback of these affinity purifica-tions is that the purification is dependent on a specific modification, and thus possible effects of other modifications and multiple modifications in factor association would not be identified. In this regard, our unbiased approach of purifying tailassociated proteins directly from cells by expressing H3 tail domains is more suitable for identifying tail-interacting factors than these in vitro purification methods.
Our analysis to examine cellular modifications of expressed H3 tails has confirmed acetylation, methylation, and phosphorylation of the tail domains at major modification sites (Fig. 3). However, the purified proteins only showed HMT activities mainly acting on H3 (Fig. 4), indicating that acetylation and phosphorylation of ectopic H3 tails arise from transient actions of cellular histone acetyltransferases and kinases. Consistent with HMT activities of the tail-associated proteins, we identified MLL3, G9a, ASH1, and EuHMT specific for Lys-4 and Lys-9 of H3 and CARM1 specific for Arg-17 of H3. However, although the tail-associated factors can methylate Lys-27 and Lys-36, we could not detect any factors that are responsible for these modifications. Although we do not have a clear explanation, it is possible that MLL3, G9a, ASH1, and/or EuHMT associated with ectopic H3 tails can methylate Lys-27 and/or Lys-36 together with other associated factors. It is also possible that Lys-27/Lys-36-specific activities stem from unknown or other associated proteins in the purified factors.
Our results also have shown that ectopic H3 tails are stably associated with chromatin-specific proteins that are repressive (HP1, HDAC5/9, Mi-2b, G9a, and TIF1␤) and active (Spt16, SSRP1, nucleolin, HMGB1, and TRAP150) in transcription. Apart from a specific recognition of HP1 to trimethylated H3 tails, it is currently unclear if these proteins directly interact with ectopic H3 tails and if any H3 tail modification can assist the interaction of these proteins with H3 tails. We are currently investigating a contribution of H3 tail to the control of initial recruitments of these factors during transcription, possibly through its specific modification. We also note that SAP155/ 145/130 subunits of the splicing factor 3b (SF3b) are stably associated with ectopic H3 tails. These results support the possible role of H3 tail in regulation of cotranscriptional splicing processes. Indeed, recent studies identified SAP130 in GCN5Lcontaining STAGA and TFTC complexes (32,33), which recognize and acetylate H3 tails. Thus our observations bear an important implication on a possible coupling between H3 tailmediated chromatin remodeling and alternative splicing. Another important point with our results is that the H3 tailassociated proteins are devoid of bromodomain-containing factors (e.g. PCAF, TAFII250, and BRG1) that have been shown to interact with acetylated H3 tails (20). Although it needs to be determined, it may be that other modifications such as Lys-9/ Lys-27 methylation and Ser-10 phosphorylation within expressed H3 tails inhibit stable association of bromodomains with acetylated H3 tails; thus it would be of considerable interest to examine whether blocking of these neighboring modifications stimulates the interaction between acetylated H3 tails and bromodomain-containing factors in future studies.
Consistent with the presence of multiple chromatin-specific activities in the H3 tail-associated proteins, our transcription assays showed that the purified proteins significantly up-regu-late p300-mediated chromatin transcription (Fig. 5C, lanes  13-15) but not DNA transcription. The inability of the associated proteins to enhance transcription in the absence of p300 and acetyl-CoA (Fig. 5C, lanes 16 -18) further indicates that chromatin acetylation (mediated by p300) is prerequisite for the action of the tail-associated proteins. The most likely interpretation of our transcription results is that the H3 tail-associated proteins interact with activators to selectively recognize p300-mediated acetylation at the promoter region of chromatin template. This promoter localization of the proteins will further induce relief of nucleosome-mediated repression at the promoter to facilitate the formation of preinitiation complex. Given their general linkage to Lys-9 methylation-induced heterochromatic gene silencing, it is unexpected to see that the HP1-containing wild type tail-associated proteins could activate chromatin transcription. However, it is notable that we only study the effect of the tail-interacting proteins on acetylation-mediated transcription, and thus HP1 proteins associated with ectopic H3 tails will have a minimal effect in transcription. Thus further characterization of the H3 tail-associated factors with various histone modifying cofactors such as HMT and histone kinase will facilitate our understanding of the molecular details for transcription regulation by the H3 tail.
Lys-9 and Lys-27 Methylation as a Key Determinant of H3 Tail-Factor Interactions-As a first step to delineate the contribution of H3 tail modifications in association of specific factors, the effect of Lys-9 and Lys-27 methylation was investigated by their mutations to arginine. In contrast to methylations of Lys-4 and Lys-36 that are associated with active chromatin, these methylations are known to induce chromatin condensation for transcription repression (17). Unexpectedly, our experiments revealed that Lys-9 and Lys-27 mutations significantly inhibited Lys-4 dimethylation and Lys-36 monomethylation (Fig.  3D) of the ectopic H3 tails, but not Lys-4/Lys-36-specific HMT activities of the tail-associated factors (Fig. 4C, compare lane 5 and lane 6). Although we do not have a clear explanation for these results, it is tempting to speculate that the methylations of ectopic H3 tails at Lys-9 and/or Lys-27 (especially di-/trimethylations at Lys-9 and/or monomethylation at Lys-27) in vivo could either directly or indirectly modulate HMT activities specific for Lys-4 and Lys-36. This dependence will allow cells to keep balance between active methylations at Lys-4 and Lys-36 and repressive methylations at Lys-9 and Lys-27.
Our mass spectrometry analysis also indicated that proteins associated with wild type H3 tail are different from those associated with Lys-9/Lys-27-mutated H3 tail, supporting specific role of methylations at these two sites in factor recruitment. A major finding is that these two marks are specifically required for association of repressive factors, including HP1, HDAC5/9, Mi-2b, G9a, and TIF1␤. We did not attempt to delineate precisely whether blocking Lys-9 methylation or Lys-27 methylation is a major cause for dissociation of these factors from the ectopic H3 tail. However, based on the pioneering works that established specific interaction between the HP1 chromodomain and Lys-9-methylated H3 tail and the recent model that describes HP1-mediated recruitment of multiple repressive factors (45,50), we assume that Lys-9 mutation of ectopic H3 tail is the major cause for the factor dissociation. It is also pos-sible that the mutation of lysine substrates itself would somehow affect the association of factors with ectopic H3 tails. However, the dissociation of only repressive factors by Lys-9 and Lys-27 mutations strongly argues against this possibility.
Significantly, the mutant H3 tail-associated proteins showed activity similar to that of the wild type tail-associated proteins in our transcription assays. Because we only checked effects of the proteins on acetylation-mediated transcription, it will be of interest to check if the wild type and mutant tail-associated proteins could differentially regulate transcription when chromatin templates are acetylated by p300 and methylated by H3-K9/K27-specific HMTs. It is likely that the H3-K9/K27 methylation of chromatin might conceivably result in the recruitment of repressive factors (HP1, HDAC5/9, Mi-2b, G9a and/or TIF1␤) from the wild type H3 tail-associated proteins and consequent repression of p300-mediated transcription. However, because all repressive factors specifically recognizing H3-K9/K27 methylation are dissociated in the mutant tail-associated proteins, we expect to find H3-K9/K27 methylation of chromatin to minimally change the effect of the mutant tailassociated proteins in p300-mediated transcription. Although more work is required to clearly understand the details, the results presented in this study are sufficient to demonstrate a functional complexity of H3 tail-mediated transactivation accompanying the multiple modifications and recruitment processes. Application of differentially mutated H3 tails in the described purification will undoubtedly facilitate our efforts to gain a full understanding of the mechanism of action of the H3 tail and its modification in transcription.